"...although man did not cause variability and cannot prevent it, he can select, preserve, and accumulate the variations given him by the hand of nature almost any way he chooses; and certainly produce a great result...."Although he did not understand how diversity arises, Darwin understood that we can alter germplasm as needed by preserving and accumulating the variations found in nature. These are the major goals of any germplasm program.
"...domestic races of plants often exhibit an abnormal character, as compared with natural species; for they have been modified not for their own benefit, but for that of man."
A diverse germplasm is necessary for breeding both established and new crops. Thus, the improvement of new industrial crops is not appreciably different from the enhancement and breeding of more established food, feed, and fiber crops (Thompson 1990a; White et al. 1994). In both, the plant breeder takes the extant germplasm and searches for genetic variability in desired traits. The major differences are that new crops plant breeders often are working with a species that is not as yet domesticated, they are unfamiliar with the species, and are starting with a limited and frequently un-evaluated or exotic germplasm base (Thompson 1990a).
In conventional crops the barriers to domestication were overcome thousands of years ago. Yet access to a diverse germplasm is still essential for the enhancement of specific traits (e.g. a specific disease resistance) in already well adapted and high yielding lines. In new crops a diverse germplasm may be even more important, since genetic diversity must be available in order to overcome the barriers to domestication, as well as, increase yield and produce lines adaptable to cultivation in diverse environments (Thompson et al. 1992). Yet, since little work has been done on these species, often relatively few accessions have been collected and are available to plant breeders.
We identify five major elements in new crops research and development programs, that encompass all aspects from domestication to commercialization. These are (1) germplasm collection, (2) germplasm evaluation, (3) germplasm enhancement and development, (4) cultivar development and evaluation, and (5) commercialization (Thompson 1990a). The first three are concerned with germplasm use and are the emphasis of this paper, in which we examine three potential new crops (vernonia, lesquerella, and guayule) for use in arid regions. The fourth step involves the utilization of improved germplasm in cultivar development for specific environments and cultural systems, and the fifth is the acceptance by industry and consumers and scale-up of production. The last two steps will not be discussed in this paper.
Germplasm collection should involve an understanding of the taxonomy and ecology of the prospective crop plant, so the collector can identify potential sources of genetic variability. After seed and plants are collected and documented, the germplasm is increased and must be maintained and stored under proper conditions before distribution. Established crops have the advantage of a long history of collecting, and their germplasm base has a higher probability of reflecting the natural genetic diversity. New crop germplasm is often unmethodical in its collection either because it was available while a collector was in the region or collected deterministically, seeking a specific trait. The problem with deterministic collecting is that breakthroughs often occur in hybrids of very divergent genetic stocks, in which neither line contained the desired characteristic (Knapp 1993; Ray 1994; Thompson et al. 1994b).
Germplasm evaluation is critical in new crop development. Rarely can a plant species be taken from the wild and converted to a commercial crop without considerable manipulation of their genetic resources. In germplasm evaluation the barriers to domestication must be identified early and the germplasm searched for genetic diversity by which to overcome these obstacles. These barriers might include adaptation, growth habit, pollination and reproduction, mode of propagation and distribution (which include seed shattering and dormancy), and pest and stress tolerance. After the barriers to domestication are identified, initial productivity and yield estimates can be made, potential chemical and nutritional utilization evaluated, and perhaps even an initial economic assessment.
In germplasm enhancement and development the barriers to domestication are addressed through plant breeding. This begins with the identification, isolation, and selection of desired traits from the available germplasm. These genes, assuming the differences are genetic, are transferred by whatever means available. These techniques might be genetic or cytogenetic, including intraspecific and interspecific hybridization, tissue culture, embryo culture, protoplast fusion, and possibly even genetic engineering technology, although this is doubtful at such an early stage in crop development. Once improved lines are developed, they should be released so that they are available to other breeders, and can become the basis upon which cultivars are developed.
In 1954, Gunstone discovered that seed oil of Vernonia anthelmintica contained high amounts of epoxy fatty acids. However, breeding and cultural work stopped when essentially no genetic variability was identified for seed retention. Perdue at al. (1986) reported the finding of V. galamensis subsp. galamensis var. ethiopica, which had high oil content and good seed retention. The problem was that var. ethiopica is a short-day plant and does not flower in the continental United States (Dierig and Thompson 1993).
In order to achieve full domestication of vernonia for cultivation and production in the United States, several barriers must be overcome. The desired traits are day-neutrality, autofertility, non-dormant seed germination, good seed retention, increased uniformity of seed maturity, and high oil and vernolic acid contents (Dierig and Thompson 1993). The first major breakthrough came when it was found that some V. galamensis subsp. galamensis var. petitiana germplasm was day-neutral (Thompson et al. 1992). V. galamensis subsp. galamensis var. petitiana was grown at six locations across the United States and highly significant differences were found in oil content (34.5% to 44.3%), vernolic acid (61.0% to 80.0%), and seed weight (1.87 to 2.92 g/1000) (Thompson et al. 1994a).
Dierig and Thompson (1993) found while evaluating germplasm, that var. petitiana was self-sterile, allowing for ease of crossing with var. ethiopica. Crosses were made forming intraspecies hybrids, and selections were performed for self-fertility, day-neutral photoperiodic response, good seed retention, non-dormant seed, and high yield (both for seed and oil) (Thompson et al. 1994b). The top 20% of the selections were grown in a winter nursery in 1993, and this seed was grown at 10 locations across the United States in 1994.
The intraspecific hybrids have been very successful, but only a limited number of accessions have been used in the crosses to date (Thompson et al. 1994b). Only 30 plants from one accession of var. petitiana were used as seed parents in the formation of hybrids. Pollen parents for all hybrids consisted of one accession (18 plants) of var. ethiopica, two accessions of var. galamensis (3 and 4 plants, respectively), one accession (3 plants) of a variety that has yet to be determined, and finally one plant from one accession of subsp. mutomoensis. This narrow genetic base is a product of the limited germplasm, since in the National Plant Germplasm System (NPGS) there are only five accessions of var. petitiana, one accession of var. ethiopica, and four accessions of var. galamensis (White et al. 1994). The danger here is that the genetic base will be narrowed too quickly while making selections in only a few environments.
Lesquerella fendleri has been targeted as the primary species for domestication, with variation in lesquerolic acid from 55% to 60%. Seed yields of 1,400 kg/ha have been obtained from an essentially unselected population (Thompson et al. 1992), and through selections seed yields have increased to 1,700 kg/ha, with 30% oil and 59% lesquerolic acid. To date, all of the selected families used in cultural and processing research have a very narrow genetic base (Thompson and Dierig 1994).
Early selections were primarily to improve plant height, vigor, and seed yield. Now the breeding program is looking for autofertility, although there appears to be limited variation in the available germplasm, determinate growth habit, and early flowering. Up to 15% of L. fendleri plants in seed production populations have been found to be male sterile (Thompson and Dierig 1994). The inheritance of male-sterility is being investigated and is thought to have real potential for use in crosses to increase the genetic diversity in breeding populations.
The need for acquisition and utilization of new sources of lesquerella germplasm was recognized and addressed by workers at the U.S. Water Conservation Laboratory in Phoenix, Arizona. Collections were made in 1993 and 1994. In 1993, 44 L. fendleri accessions, 38 accessions from nine other Lesquerella species, and two different Physaria species were added to the working germplasm collection at Phoenix, Arizona. In 1994, another 41 L. fendleri accessions and 51 other accessions from 10 other Lesquerella species were added to the collection (Dierig et al. 1996). Evaluation of this germplasm is presently taking place, and new variation for seed weight and growth habits have been found. Selecting for autofertility is still a high priority.
Unlike in Hevea, where latex flows in continuous ducts, in guayule the whole plant is harvested, ground, and the latex extracted chemically (Thompson and Ray 1989). Harvest of native stands and initial use of guayule as a source of natural rubber began in the late 1800s and became a major source for the United States and Mexico in the early 1900s. Unfortunately, native stands were rapidly depleted and much genetic diversity was undoubtedly lost.
Most guayule germplasm consists of apomictically reproducing triploids (3n = 54) and tetraploids (4n = 72), which have received the most attention in the breeding programs. Sexually reproducing, largely self-incompatible diploids (2n = 36) are found in a very restricted area in Mexico. Most related species are also diploid, although some have a polyploid series (Thompson and Ray 1989). A recent germplasm collection of new diploids, and their use in a recurrent selection program, along with use of interspecific hybridization is adding dimension to the total breeding effort (Estilai and Ray 1991).
Most of the current germplasm being utilized in breeding programs originated from material developed from two major collections during the Emergency Rubber Project. Most of this material traces back to a small number of accessions collected in a very limited area in the Mexican state of Durango. Twenty-two of the original 26 lines in the National Plant Germplasm System (NPGS) are from Durango. In fact, 15 of the accessions are descended from a bulk seed collection of only five plants at one location, and the sexual diploid accession was also from only five plants at one location.
Surprisingly a large amount of variability for rubber and resin quantity and quality and plant growth characteristics have been shown to exist within the apomictically reproducing polyploid germplasm (Ray et al. 1990). The facultative nature of apomixis in the polyploid material apparently serves as a mechanism for both conservation and propagation of a wide array of genetic and chromosomal variation. Many new apomictic single plant and line selections have been made with significantly increased yields (Thompson and Ray 1989; Estilai and Ray 1991).
Today there are 29 guayule accessions that have PI numbers in the NPGS. Two of these are interspecific hybrids and all are polyploid except two diploid accessions. Another 158 selections have been entered in the Germplasm Resources Information Network (GRIN) system, 111 of which are guayule. Much additional germplasm resides in the working collections in Arizona, California, and Texas. There may also be as many as 3,000 accessions collected from 310 locations in six Mexican states at the Universidad Autonoma Agraia in Saltillo, Coahuila (Thompson et al. 1992).
Selection has been primarily for rubber yield, with some work in disease resistance, stress tolerance, and multiple uses. Significant gains have been made in the sexual diploid populations, however, diploid lines generally yield lower than polyploid lines. The facultative nature of apomixis in polyploid guayule has been essentially ignored in the breeding programs. Breeding within polyploids has generally been by either mass selection or half-sib families of single-plant selections. Since apomixis is facultative, it is difficult to identify plants that are genetically different, and once improved lines are obtained, keeping them pure breeding.
Yields have been significantly increased (Thompson and Ray 1989; Estilai and Ray 1991), now the emphasis should turn to the barriers keeping guayule from becoming fully domesticated. The primary barrier is apomictic reproduction, and emphasis should be upon aggressively selecting separately for both obligate sexual and apomictic reproduction. Several barriers are concerned with seed production, and include dormancy, viability, genetic uniformity, and uniform maturity. Additionally, shortening the length of time until harvest should be seriously considered.
Progress in breeding programs is scale dependent, and new crops programs have few individuals working on domestication and improvement. Unfortunately, there are often very short time frames during which new crops have a real chance of industrial acceptance. Because of these short time frames, and the number of plant breeders involved, new crops germplasm is often used directly without major improvements. These germplasm lines are really little more than landraces, thus it is very hard to anticipate yields and needed cultural practices in different environments.